You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

140 lines
3.8 KiB
Plaintext

TITLE:: FluidMDS
summary:: Dimensionality Reduction with Multidimensional Scaling
categories:: Dimensionality Reduction, Data Processing
related:: Classes/FluidMDS, Classes/FluidDataSet
DESCRIPTION::
https://scikit-learn.org/stable/modules/manifold.html#multi-dimensional-scaling-mds
CLASSMETHODS::
METHOD:: new
Make a new instance
ARGUMENT:: server
The server on which to run this model
METHOD:: euclidean
Euclidean distance (default)
METHOD:: sqeuclidean
Squared Euclidean distance
METHOD:: manhattan
Manhattan distance
METHOD:: max
Minowski max
METHOD:: min
Minowski max
METHOD:: kl
Symmetric Kulback Leiber divergance (only makes sense with non-negative data)
METHOD:: cosine
Cosine distance
INSTANCEMETHODS::
PRIVATE:: init
METHOD:: fitTransform
Fit the model to a link::Classes/FluidDataSet:: and write the new projected data to a destination FluidDataSet.
ARGUMENT:: sourceDataset
Source data, or the dataset name
ARGUMENT:: destDataset
Destination data, or the dataset name
ARGUMENT:: k
The number of dimensions to reduce to
ARGUMENT:: dist
The distance metric to use (integer, 0-6, see flags above)
ARGUMENT:: action
Run when done
EXAMPLES::
code::
//Preliminaries: we want some audio, a couple of FluidDataSets, some Buffers, a FluidStandardize and a FluidMDS
(
~audiofile = File.realpath(FluidBufPitch.class.filenameSymbol).dirname +/+ "../AudioFiles/Tremblay-ASWINE-ScratchySynth-M.wav";
~raw = FluidDataSet(s,\mds_help_12D);
~reduced = FluidDataSet(s,\mds_help_2D);
~audio = Buffer.read(s,~audiofile);
~mfcc_feature = Buffer.new(s);
~stats = Buffer.new(s);
~datapoint = Buffer.alloc(s,12);
~standardizer = FluidStandardize(s);
~mds = FluidMDS(s);
)
// Do a mfcc analysis on the audio, which gives us 13 points, and we'll throw the 0th away
// Divide the time series in to 100, and take the mean of each segment and add this as a point to
// the 'raw' FluidDataSet
(
~raw.clear;
~norm.clear;
FluidBufMFCC.process(s,~audio,features:~mfcc_feature,action:{
"MFCC analysis.complete. Doing stats".postln;
fork{
var chunkLen = (~mfcc_feature.numFrames / 100).asInteger;
100.do{ |i|
s.sync; FluidBufStats.process(s,~mfcc_feature,startFrame:i*chunkLen,numFrames:chunkLen,startChan:1, stats:~stats, action:{
~stats.loadToFloatArray(action:{ |statsdata|
[statsdata[0],statsdata[1]].postln;
~datapoint.setn(0,[statsdata[0],statsdata[1]]);
s.sync;
("Adding point" ++ i).postln;
~raw.addPoint(i,~datapoint);
})
});
if(i == 99) {"Analysis done, dataset ready".postln}
}
}
});
)
//First standardize our dataset, so that the MFCC dimensions are on comensurate scales
//Then apply the MDS in-place on the standardized data to get 2 dimensions, using a Euclidean distance metric
//Download the dataset contents into an array for plotting
(
~standardizer.fit(~raw);
~standardizer.transform(~raw, ~reduced);
~mds.fitTransform(~raw,~reduced,2, FluidMDS.euclidean);
~reducedarray= Array.new(100);
fork{
100.do{|i|
~reduced.getPoint(i,~datapoint,{
~datapoint.loadToFloatArray(action:{|a| ~reducedarray.add(Array.newFrom(a))})
});
s.sync;
if(i==99){"Data downloaded".postln};
}
}
)
//Visualise the 2D projection of our original 12D data
(
d = ~reducedarray.flatten(1).unlace.deepCollect(1, { |x| x.normalize});
// d = [20.collect{1.0.rand}, 20.collect{1.0.rand}];
w = Window("scatter", Rect(128, 64, 200, 200));
w.drawFunc = {
Pen.use {
d[0].size.do{|i|
var x = (d[0][i]*200);
var y = (d[1][i]*200);
var r = Rect(x,y,5,5);
Pen.fillColor = Color.blue;
Pen.fillOval(r);
}
}
};
w.refresh;
w.front;
)
::